Title: Using Eye Gaze to Play Mind Card-game using Neural Network

Year of Publication: Aug - 2019
Page Numbers: 54-60
Authors: Qingyao Hu, Seanglidet Yean, Jigang Liu, Bu Sung Lee, Deepu Rajan, Romphet Phattharaphon
Conference Name: The Fifth International Conference on Electronics and Software Science (ICESS2019)
- Japan


Since the invention of smartphones, touch has always been the primary interface for humans to interact with smartphones and tablets. As technology advances, researchers have been exploring other forms of interfaces such as voice control, virtual reality, handwriting digitizers, etc. One of the least explored and potential human-computer interfaces is eye gaze tracking which is the method of measuring the point of gaze. Studies of eye movement and eye tracking has been attracting interests from researchers and has progressed from eye-attached tracking to optical tracking and currently machine learning based tracking. In this paper, we proposed two ResNet-based convolutional neural network (CNN) models for card prediction which is a classification problem. Eye gazes are categorized into 32 areas whereby each eye gaze corresponds to an area on the mobile device’s display. The experimental results show that we can achieve accuracy as high as 81%.